In agricultural systems, multiple objectives and uncertainty are often part of the game when optimization is at stake. Multiobjective dominance rules cannot be simply applied due to this uncertain behavior. We propose some extensions of the well-known Pareto rules to enable the discrimination of multicriteria dominating groups of outcomes. These groups are either the various uncertain outcomes of a decision, or more generally a set of outcomes associated to different decisions and/or different random occurrences. Based on the Pareto dominance rules, we propose definitive, acceptable and undecidable dominance comparisons with regard to two candidate groups. The comparisons of all candidate groups allow to rank them from a multicriteria evaluation perspective. This ranking process is used as the evaluation step of a hierarchical decomposition procedure where the best ranked region is selected as the one to be investigated further. We apply these multicriteria extensions to look for optimal irrigation strategies. The yield, the total amount of water and the number of irrigation rounds are simulated to get economical, environmental and social perspectives simultaneously. Although the computation requires a high amount of simulation runs, the algorithm succeeds in reproducing the front of the non-dominated evaluations. The major interest resides in the width of the front achieved. This new information gives direct indication to the decision maker about the reliability of the outcomes with regard to the weather uncertainty, as well as the sensitivity of the outcomes with regards to the strategies application.